Death of the keyword and the (possible) future of semantic search
Semantic search has generated a lot of buzz this summer, which has search marketers considering whether the future of optimization will be focused on content or keywords.
So what is the big appeal about semantic search? Take into account this scenario: Suppose you're looking for a new luxury car. You type the phrase “best luxury car” into the search engine of your choice, which results in numerous articles, consumer reports, user reviews, etc. -- all aiming to influence your decision. However, the results you really receive are for any page that has text containing “best,” “luxury,” and “car.”
Typical search engines will show you “how to choose” articles, while Web sites for Mercedes-Benz, BMW, or Audi are not listed, even though they may be relevant for the search you've performed.
Semantic search attempts to further refine your results by not just offering the generic Mercedes-Benz homepage, but directing you to the PDF fact sheet or some other relevant material that may help you make your decision or purchase.
Search as we know it today is also “blind” to the context around what you're searching for and why. For instance, if you are looking to go on a cruise and search for the term “cruise” on Google, you get results for Carnival cruises, along with Tom Cruise. In other words, today's search just isn't that good at figuring out exactly what you want based on your searches. Figuring that out is the key to targeting the proper audience with your advertising thereby driving better business (ever wonder why the engines are so insistent upon relevancy?).
To a certain degree, the major search engines are already attempting to add elements of semantic search. On Google, the more you search for a certain phrase without clicking on a result, the more you'll begin to receive alternative ads. This is because the engine assumes that the initial results displayed were not relevant to you.
Semantic search calls for the search engines (and their designers) to have a greater understanding of what the searcher is actually trying to do (buy, sell, learn); and that's no simple task when searches typically have only 1 - 3 terms.
Some engines such as Powerset (recently purchased by Microsoft) and Hakia are now making a full throttle attempt at semantic search, allowing users to search by posing questions such as “What are the benefits of owning a home?” These engines just haven't caught on yet, because in many cases results on the traditional search engines are still more relevant. A search on Hakia for “laptop computer” brings up results for reviews, comparison sites and retail sites, very similar to the diverse set of results Google has for the same query.
Since semantic search is clearly growing, albeit slowly, what does the future hold?
Imagine walking up to your computer, speaking to a search engine and getting results based on the inflection of your voice. For example, if you say sharply “credit card!” the search engine will know you're upset with your current provider based on your inflection and show you competitors, rather than information on how to choose a credit card or finding the best credit card for you.
Traditional search engines will likely continue to dominate the marketplace for at least the next decade. For semantic search to become successful, technology will need to catch up to imagination.